Calibration in both x and y direction

classic Classic list List threaded Threaded
6 messages Options
duan duan
Reply | Threaded
Open this post in threaded view
|

Calibration in both x and y direction

Dear Mr. William,

Thank you very much for developing and sharing the PIVlab.

I have one question about the calibration for using the PIVlab.

In PIVlab, the calibration function converts the distance in px to real distance in mm, where the distance dist=sqrt((xposition(1)-xposition(2))^2 + (yposition(1)-yposition(2))^2);.

However the image distortion might be different in x and y direction.

If I want to do the calibration in both x and y direction: one calibration factor in x and one calibration factor in y direction, which functions in which m files should be modified?

I think I might have to modify the function "calccali" and separate the "calxy" into "calx" + "caly" and the others...

Could you make a little guideline for this?

Looking forward to hearing from you.

Regards,

duan
William William
Reply | Threaded
Open this post in threaded view
|

Re: Calibration in both x and y direction

Hi, yes you prabably need to do it this way. You might however also consider to undistort the images before loading them into PIVlab, that might be easier.

Please consider donating for the PIVlab project:
http://paypal.me/pools/c/81Y8wBVi4y

The developer needs to update to a recent Matlab release in order to keep the project running.

duan duan
Reply | Threaded
Open this post in threaded view
|

Re: Calibration in both x and y direction

Dear William,

Thank you very much for the quick reply.

Could I ask one more question, :D

Do you have any recommendation about the  techniques or softwares for image pre-processing?

I am still looking for a tool to do the image pre-processing, for example to subtract the background image and now to undistort the image.

Thank you so much for your attention.

Duan



On 27 May 2015 at 15:31, William [via PIVlab - Time-Resolved Digital Particle Image Velocimetry Tool for MATLAB] <[hidden email]> wrote:
Hi, yes you prabably need to do it this way. You might however also consider to undistort the images before loading them into PIVlab, that might be easier.


To unsubscribe from Calibration in both x and y direction, click here.
NAML



--
Qingfeng Duan
COPPE/UFRJ
Universidade Federal do Rio de Janeiro
Tel:+55 (21) 7938 1582
William William
Reply | Threaded
Open this post in threaded view
|

Re: Calibration in both x and y direction

Hi,
I don't ahve any recommendations for this... Probably it's best to write your own code in Matlab to do the pre-processing. There is also a camera calibration toolbox that should be able to undistort your images.

Please consider donating for the PIVlab project:
http://paypal.me/pools/c/81Y8wBVi4y

The developer needs to update to a recent Matlab release in order to keep the project running.

duan duan
Reply | Threaded
Open this post in threaded view
|

Re: Calibration in both x and y direction

Dear William,

I will try the calibration toolbox.

Thank you so much.

Duan

On 27 May 2015 at 16:31, William [via PIVlab - Time-Resolved Digital Particle Image Velocimetry Tool for MATLAB] <[hidden email]> wrote:
Hi,
I don't ahve any recommendations for this... Probably it's best to write your own code in Matlab to do the pre-processing. There is also a camera calibration toolbox that should be able to undistort your images.


To unsubscribe from Calibration in both x and y direction, click here.
NAML



--
Qingfeng Duan
COPPE/UFRJ
Universidade Federal do Rio de Janeiro
Tel:+55 (21) 7938 1582
Tobi Tobi
Reply | Threaded
Open this post in threaded view
|

Re: Calibration in both x and y direction

Undistorting pictures is pretty easy with Irfanview's batch processing: in irfanview type "b"; in the pop-up choose "batch-conversion" and mark "special options"; click on the button next to it, choose "change size" and make your changes. Now choose the pictures you need to undistort (add them to the queue) and hit "start".
Background deletion is unfortunately more complex. I cant see a different way than using MATLAB for that.